Research And Improvement Of Multi-keyword Searchable Encryption Schemes Supporting Multi-user Access

Cloud computing has strong computing power and storage capacity,and relies on sharing of resources to deliver convenient and flexible services to enterprises or users,so that they can maximize their own interests with minimal cost.Although cloud storage provides convenience to data storage and usage,sensitive data stored on the cloud servers may bring a series of privacy problems.Encrypting data before uploading it to the cloud is the general way to protect the user's privacy from leakage.The traditional way of downloading all the data and decrypting locally is clearly impractical,due to the high computation cost.In order to solve the above problem,searchable encryption(SE)is proposed,which can search over the encrypted data and has less computing overhead.However,most existing SE schemes only support single-keyword search and single-user scenario.First,due to single-keyword search,the user may download a large amount of returned files in database.Second,limited to a single-user scenario and it is possible that anyone can pretend to be data owner to upload forgery data to CSP.To solve above problems of SE,the main work of the thesis are summarized as follows:1)We proposed a multi-user and multi-keyword searchable encryption supporting anti-forgery(AMMSE)scheme.AMMSE uses an inverted index structure to construct an encrypted index to support efficient multi-keyword search,and uses ciphertext policy attribute-base encryption(CP-ABE)to encrypt the search key to reduce interaction between the data owner and the user.In addition,we embed a specific secret in encrypted index,so that only data owner instead of any user uploads data.Security analysis shows that our scheme is secure under non-adaptive attacks.The experimental results show that the encryption efficiency of AMMSE is practical than other related schemes.2)In order to achieve fast accurate search and improve the user's search experience,we propose an efficient multi-keyword ranked search supporting CP-ABE test(MRST)scheme.We introduce the TF-IDF model,numeric hierarchy structure,and index array to achieve efficient encryption and ranked search,and use secure KNN to ensure data privacy in indexand query encryption.We also improve CP-ABE to supporting test that ensure the user can decrypt all returned files.The security analysis shows that MRST is secure,and the experimental results show that MRST is efficient than other comparison schemes,and can improve the user's search experience.